集成学习算法的研究与应用

Study of ensemble algorithm and its application

集成学习算法的思想就是集成多个学习器,并组合它们的预测结果,以形成最终的结论。典型的学习模型组合方法有投票法,专家混合方法,堆叠泛化法与级联法,但这些方法的性能都有待进一步提高。提出了一种新颖的集成学习算法--增强的集成学习算法(ReinforcedEnsemble)。ReinforcedEnsemble集成算法由两大部分组成:ReinforcedEnsemble特征提取算法与ReinforcedEnsemble基分类器。通过实验,将ReinforcedEnsemble算法与其他集成学习算法进行了性能比较。实验结果表明,所提出的算法在多项指标上均达到最优。

The idea of ensemble learning is to employ multiple learners and combine their predictions. The typical methods of combining multiple models such as bagging, boosting, stacking error correcting output codes, voting, mixtures of experts, stacked generalization and cascading. Though a considerable effort has been put into developing statistical models and algorithmic strategies for classification, the accurate of the classification has been proven to be very challenging. A novel ensemble algorithm, ReinforcedEnsemble is proposed. ReinforcedEnsemble ensemble algorithm consists of two parts, ReinforcedEnsemble feature extraction algoritlun and ReinforcedEnsemble base classifier. The performance between ReinforcedEnsemble and other ensemble algorithm in the experiments is compared. The experimental results show that the proposed algorithm is optimal in a number of indicators.